---
title: "Howard Marks of Oaktree makes 180-degree turn on AI after Claude tutorial. Here's how he suggests investors approach it."
type: "News"
locale: "en"
url: "https://longbridge.com/en/news/277192641.md"
description: "Howard Marks, co-founder of Oaktree Capital Management, has shifted his stance on AI after a tutorial with Anthropic's Claude AI model. Previously skeptical, he now acknowledges AI's revolutionary potential but warns against being overly optimistic about investments. Marks emphasizes a balanced approach, advising against going 'all-in' or 'all-out' on AI. He recognizes AI's ability to make intelligent decisions and its implications for investment management, while also expressing concerns about labor displacement and the risks associated with certain AI investments."
datetime: "2026-02-27T11:52:40.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/277192641.md)
  - [en](https://longbridge.com/en/news/277192641.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/277192641.md)
---

> Supported Languages: [简体中文](https://longbridge.com/zh-CN/news/277192641.md) | [繁體中文](https://longbridge.com/zh-HK/news/277192641.md)


# Howard Marks of Oaktree makes 180-degree turn on AI after Claude tutorial. Here's how he suggests investors approach it.

By Jules Rimmer

No one should be 'all-in' or 'all-out' of AI, says Oaktree co-founder

Howard Marks is much less skeptical about AI than he was just a few months ago.

In December, Howard Marks published an investment memo titled, "Is it a bubble?" that expressed some of his skepticism and reservations about artificial intelligence and the stock-market boom it had created.

A few months and one tutorial from Anthropic's Claude AI model later, Marks's latest investment memo has the header "AI hurtles ahead." It's fair to say that if Marks hasn't been completely converted to a bull on AI investment, he is now convinced of the revolutionary technological benefits it brings. He guesses its potential is underestimated but warns that may not mean the investment price is right.

As the co-founder and co-chairman of Oaktree Capital Management, Marks's opinions are well-regarded. So when a few months back he cautioned about "the regularity of bubbles," "bubble behavior," and that the "end result is inevitably painful," investors in AI stocks sat up and listened.

It's not quite a Damascene conversion, but he's certainly much more open to the technology's upside potential.

Encouraged by some of those investors to try out a few of the AI models available, Marks placed the Claude bot under scrutiny and came away thoroughly impressed. It's not quite a Damascene conversion, but he's certainly much more open to the technology's upside potential. For example, he kicks off, "Before I start in, I want to try to communicate the level of awe with which I viewed Claude's output."

Given the week of bad PR that AI has endured in the aftermath of a viral Citrini Research paper, whose portrait of economic doom was heightened by the Cash App and Square company Block's decision to halve its workforce and shed 4,000 jobs, the AI industry needed a boost. Now it has received what it can plausibly describe as a ringing endorsement from one of America's most reputable and experienced investors.

What impressed Marks was the ability of AI to act autonomously, he said. "It was making intelligent decisions. It had something that felt, for the first time, like judgment. Like taste," he says.

Commenting on the massive leap in capabilities in just a few months, Marks makes the point that in terms of analyzing information: "AI possesses a lot of the qualities one needs to be a good investor." Citing ChatGPT-5.3 Codex and Claude's Opus 4.6, Marks notes AI is making "intelligent decisions and even contributing to its own code and improvement, creating an accelerating feedback loop."

When it comes to making investment decisions, he now acknowledges that, as indexation became a competitive threat to active investment management, AI may well push out "average" investors, partly because its rationale won't be tainted with human emotions like greed and fear.

Weighing the scales now, as opposed to his December debate over whether the AI investment theme is a bubble or not, Marks concludes by emphasizing that just because the technology is real and revolutionary doesn't mean that every investment in it is safe.

He regards circular revenue and "lottery-ticket startups" as risky. He is deeply worried about "labor displacement" but overall concludes it's just too early to say whether AI is "a fad or an illusion."

Forecasts suggest Howard Marks (and many other commentators) are right to be concerned about labor displacement.

Not knowing whether the investment in AI infrastructure can produce an adequate return, Marks finds it hard to say whether hyperscalers are undervalued or overvalued. The potential for AI is most likely underestimated, he believes, "but that's not the same as saying AI investments are on the bargain counter."

"No one should go all-in," he says, but, "by the same token, no one should stay all-out."

\-Jules Rimmer

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

02-27-26 0652ET

### Related Stocks

- [BROOKFIELD OAKTREE HOLDINGS LLC 6.625% PREFERRED UNITS SERIES A (OAK-A.US)](https://longbridge.com/en/quote/OAK-A.US.md)
- [iShares Global Tech ETF (IXN.US)](https://longbridge.com/en/quote/IXN.US.md)
- [BROOKFIELD OAKTREE HOLDINGS LLC 6.550% PRFD UNITS SERIES B (OAK-B.US)](https://longbridge.com/en/quote/OAK-B.US.md)

## Related News & Research

- [Claude Code leak: Anthropic cites human error, works to limit damage](https://longbridge.com/en/news/281408775.md)
- [Anthropic struggling with Chinese competition, its own safety obsession](https://longbridge.com/en/news/280880137.md)
- [The Next Big AI Winner Might Not Be a Tech Company](https://longbridge.com/en/news/281689441.md)
- [Napster is Evolving in the AI Era](https://longbridge.com/en/news/281749361.md)
- [Anthropic admits Claude Code users hitting usage limits 'way faster than expected'](https://longbridge.com/en/news/281229875.md)